- 19
- 3
- 1
- 1
- 1
DC
David Causeur
25
Documents
Identifiants chercheurs
- david-causeur
- 0000-0001-6910-9440
- Google Scholar : https://scholar.google.com/citations?user=kCws61IAAAAJ&hl=fr
Présentation
Research topics
===============
My research interests are in statistical methodology for various issues motivated by biological applications. More particularly, the focus of my recent papers is on the handling of dependence in high-dimensional statistical inference.
**Statistical genomics**
Dependence within high-dimensional gene expression profiles generates a strong instability of gene selection in large scale significance analysis. A proper handling of this dependence by latent factor models or more general whitening techniques improves stability of multiple testing procedures and power (see for example Friguet *et al*, 2009 \[JASA\], Friguet and Causeur, 2010 \[CSDA\], Causeur *et al*, 2011 \[JSS\], Hornung *et al*, 2016 \[BMC Bioinf.\], Hornung *et al*, 2017 \[Bioinf.\], Hébert *et al*, 2021 \[CSDA\]).
**Functional data analysis**
Functional data are discretized observations of curves. Such data are generated by various technologies such as spectroscopy or electroencephalography (EEG). More and more study designs, such as Event-Related Potentials studies in neuroscience, aim at assessing the relationship between functional data and experimental covariates. Both for signal detection (global testing) and signal identification (search for significant intervals), dependence handling strategies can be designed to be efficient for a given pattern of association signal (see for example Causeur *et al*, 2012 \[BRM\], Sheu *et al*, 2016 \[AoAS\], Causeur *et al*, 2020 \[Biometrics\]).
**High-dimensional regression and classification modeling**
Both in genomic and functional data analysis, estimation of regression and classification models has to deal with a possibly strong dependence within high-dimensional profiles of explanatory variables. Whitening procedures can help stabilizing model selection methods and improve prediction performance (see for example Perthame *et al*, 2015 \[StatCo\], Hébert *et al*, 2021 \[Under revision\]).
**Biostatistics**
I am involved in various research projects with a diversity of partners in biology, recently for multi-omic data integration issues (see Gondret *et al*., 2017 \[BMC Gen.\], Désert *et al*., 2018 \[BMC Gen.\]), peptidomics (see Suwareh *et al*., 2021 \[Food Ch.\]), electromyographic data analysis (see Comfort *et al*., 2021 \[PACA\]), etc.
Publications
- 4
- 3
- 2
- 2
- 2
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 25
- 7
- 6
- 4
- 4
- 3
- 3
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 2
- 1
- 1
- 4
- 3
- 1
- 9
- 4
|
Variable selection for correlated data in high dimension using decorrelation methodsStatlearn: Challenging problems in statistical learning, Apr 2016, Vannes, France
Communication dans un congrès
hal-01310571v1
|
FADA: an R package for variable selection in supervised classification of strongly dependent datauseR!2014, UCLA Statistics Department; Foundation for Open Access Statistics; Los Angels R user group, Jun 2014, Los Angeles, United States
Communication dans un congrès
hal-01167304v1
|
|
|
Stabilité de la sélection de variables pour la classification de données en grande dimension45 èmes Journées de Statistique, May 2013, Toulouse, France
Communication dans un congrès
hal-00913047v1
|
|
Inferring gene networks using a sparse factor model approachStatistical learning and data science, May 2012, Florence, Italy. 12 p
Communication dans un congrès
hal-02806777v1
|
Genetic analysis of a complex trait using transcriptomic data: contribution of gene regulatory network modelingXX. Conference of the plant & animal genome (PAG), Jan 2012, San Diego, United States. 1 p
Communication dans un congrès
hal-00841103v1
|
|
|
Inferring gene networks using a sparse factor model approach, Statistical Learning and Data ScienceLearning and Data Science, May 2012, Florence (IT), Italy
Communication dans un congrès
hal-00841017v1
|
|
Large-scale significance testing of high thoroughput Data with FAMT14. Conference of the Applied Stochastic Models and Data Analysis International Society, Jun 2011, Rome, Italy
Communication dans un congrès
hal-02746752v1
|
|
Inférence sur réseaux géniques par Analyse en Facteurs42èmes Journées de Statistique, 2010, Marseille, France, France
Communication dans un congrès
inria-00494802v1
|
|
Estimation conditionnelle de la proportion d'hypothèses nulles en grande dimension41èmes Journées de Statistique, SFdS, Bordeaux, 2009, Bordeaux, France, France
Communication dans un congrès
inria-00386591v1
|
Estimation of the proportion of null p-values among dependent testsPetersburg Workshop on Simulation, Jun 2009, St Petersburg, Russia
Communication dans un congrès
hal-00461648v1
|
|
Factor Analysis for Multiple Testing: an R-package to analyze a genome-scale datasetWorkshop « Statistical advances in Genome-scale Data Analysis », May 2009, Ascona, Switzerland
Communication dans un congrès
hal-00459370v1
|
|
Conditional Fdr estimation based on a factor analytic approach of multiple testingWorkshop on Simulation, Jun 2009, St Petersburg, Russia
Communication dans un congrès
hal-00461697v1
|
|
Factor Analysis for Multiple Testing: a general approach for differential analysis of genome-scale dependent dataWorkshop on Statistical advances in Genome-scale Data Analysis, May 2009, Ascona, Switzerland
Communication dans un congrès
hal-00461632v1
|
|
Factor Analysis for Multiple Testing (FAMT) : an R package for simultaneous tests under dependence in high-dimensional dataUseR!2009. Agrocampus, Jul 2009, Rennes, France
Communication dans un congrès
hal-00461700v1
|
|
Estimation conditionnelle de la proportion d'hypotheses nulles en grande dimension41èmes Journées de Statistiques, Société Française de Statistiques, May 2009, Bordeaux, France
Communication dans un congrès
hal-00461645v1
|
|
Multiple tests for high-throughput data assuming a factor modeling of dependence40th Journées de Statistiques (Société Française de Statistiques) / joint with Société Statistique, May 2008, Ottawa, Canada
Communication dans un congrès
hal-00461705v1
|
|
Approche conditionnelle des tests multiples pour données biologiques à haut-débit7ème Journée Jeunes Chercheurs (Société Française de Biométrie). INSERM, Dec 2008, Villejuif, France
Communication dans un congrès
hal-00461708v1
|
|
Accounting for a factor structure in high-dimensional data to improve multiple testing procedures6th workshop on Statistical Methods for Post-Genomic Data. Agrocampus, Feb 2008, Rennes, France
Communication dans un congrès
hal-00461702v1
|
|
Impact of Dependence on the Stability of Model Selection in Supervised Classification for High-Throughput DataInternational Indian Statistical Association (IISA) Conference on Frontiers of Probability and Statistical Science, May 2008, Storrs, United States
Communication dans un congrès
hal-00461704v1
|
Genetic analysis of a complex trait using transcriptomic data: contribution of gene regulatory network modelingXX. Conference of the plant & animal genome (PAG), Jan 2012, San Diego, United States. 2012
Poster de conférence
hal-02806842v1
|